PROJECT SUMMARY Immunohistochemistry (IHC) is widely used to classify cancer and guide clinical management of patients. A major power of IHC is the detection of markers in the spatially resolved context of tumor tissue, including cancerous and stromal elements, individual cell types, and subcellular compartments. Quantitative Immunofluorescence (QIF) is used increasingly for IHC quantification of proteins that may serve as prognostic biomarkers or targets for specific cancer therapies. A common application of QIF is simultaneous slide-based analysis of certain protein expression using tissue microarrays (TMAs), which incorporate tumor tissues from large cohorts of patients. This technology combined with machine readers allows for high-throughput molecular profiling of tumor tissues and rapid development and validation of new prognostic and predictive cancer biomarkers. The overall goal of our proposal is to develop an analytical framework of for discovery of new IHC biomarkers to improve prognostication and drug response prediction. The new statistical methods will be tailored to use the rich and untapped information on cell-level protein expression levels. Novel QIF IHC biomarkers will be optimized as flexible indices based either on distribution functions of cell-level protein expressions or on the spatial localization of the IHC signals viewed as marks of the spatial point pattern of cancer cells. The proposed studies will yield broadly applicable methods for development of new biomarkers in cancer. The improved biomarkers would support rational recruitment of patients at elevated risk of recurrence into biomarker-driven clinical trials, with a potential to have major near-term impact on clinical research and reduce mortality from cancer.